Tomographic Lesion Identification Using Multi-Scale Image Weighting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Accurate identification of lesion regions in medical images is challenging due to similar brightness and varying three-dimensional shapes, and the size of partial images affects identification accuracy, leading to potential misidentification.
Innovation Solution
An image identification method that cuts out multiple partial images of varying sizes from a tomographic image, calculates the probability of each being a lesion region, and integrates these probabilities using a calculation that emphasizes intermediate-sized images to determine the final identification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple partial images of different sizes are used for lesion identification, then identification accuracy is improved, but calculation complexity and processing time increase
Solution Approach 1:
The tomographic image is divided into multiple partial images of different sizes (first, second, and third partial images) that all include the same position. This segmentation allows the system to analyze the same region at different scales, improving identification accuracy by capturing both fine details and broader contextual information simultaneously.
Solution Approach 2:
The invention introduces a size dimension by creating partial images with different dimensions from the same position. This multi-scale approach adds a dimensional variable to the analysis, enabling the system to evaluate lesion characteristics across multiple spatial scales and thereby improve diagnostic accuracy.
2Measurement precision
If intermediate-sized partial images are given higher weight in probability integration, then lesion detection accuracy is improved, but processing complexity increases
Solution Approach 1:
Different weights are assigned to partial images based on their size characteristics. Intermediate-sized partial images are given higher weights because they provide an optimal balance between detail and context. This local quality differentiation optimizes the contribution of each partial image to the final probability calculation, improving lesion detection accuracy.
Solution Approach 2:
The invention changes the parameter of probability weighting by size category. Instead of uniform weighting, the system adjusts the weight parameter based on partial image size, with intermediate sizes receiving higher weights. This parameter change optimizes the integration process to better reflect the diagnostic value of different scale representations.
3Reliability
If multiple probability calculations are performed and integrated, then identification reliability is improved, but processing time increases
Solution Approach 1:
The system performs preliminary probability calculations for each partial image before integration. By calculating probabilities for first, second, and third partial images separately and then integrating them with appropriate weights, the system ensures thorough analysis while organizing the computational process efficiently to manage processing time.
Solution Approach 2:
The integration process combines probability information from multiple partial images with differentiated weighting, creating a feedback mechanism where intermediate-sized images contribute more strongly to the final result. This feedback structure improves reliability by emphasizing the most diagnostically valuable scale while systematically processing all available information.
Data Source
AI summary
A computer cuts out, from a tomographic image obtained by imaging the inside of a human body, a plurality of partial images, which include the same position in the tomographic image and have different sizes, calculates a probability of each of the plurality of partial images being a region of a specified lesion, calculates an integrated value by integrating the probabilities calculated from the plurality of partial images using a calculation that increases a contribution of probabilities calculated from partial images corresponding to at least an intermediate size out of the plurality of partial images, and identifies the same position as a region of the specified lesion when the integrated value exceeds a predetermined threshold.